Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L5/00—Arrangements affording multiple use of the transmission path
  • H04L5/0001—Arrangements for dividing the transmission path
  • H04L5/0014—Three-dimensional division
  • H04L5/0016—Time-frequency-code
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L25/00—Baseband systems
  • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
  • H04L25/06—Dc level restoring means; Bias distortion correction ; Decision circuits providing symbol by symbol detection
  • H04L25/067—Dc level restoring means; Bias distortion correction ; Decision circuits providing symbol by symbol detection providing soft decisions, i.e. decisions together with an estimate of reliability
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L27/00—Modulated-carrier systems
  • H04L27/32—Carrier systems characterised by combinations of two or more of the types covered by groups H04L27/02, H04L27/10, H04L27/18 or H04L27/26
  • H04L27/34—Amplitude- and phase-modulated carrier systems, e.g. quadrature-amplitude modulated carrier systems
  • H04L27/38—Demodulator circuits; Receiver circuits

Definitions

• the present invention relates to a method and a device for soft demodulation (in English "soft demapping") in an OFDM-CDMA system (orthogonal frequency division multiplexing - code division multiple access, in English "Orthogonal Frequency Division” Multiplexing - Code Division Multiple Access "). More specifically, the object of the invention is the definition of several Likelihood Ratio Logarithms (LRV) for an OFDM-CDMA system using non-binary spreading codes.
• LDV Likelihood Ratio Logarithms
• the LRV is used by data decoding algorithms working with a so-called “soft” input, that is, not reduced to the two “hard” "0” and “1” values.
• the LRV measures the probability that the input bit of the decoder is a "0” or a "1". It consists in calculating a soft value for each of the bits inside a complex symbol modulated according to a quadrature amplitude modulation or QAM (Quadrature Amplitude Modulation) and this independently for the bits the same received QAM symbol.
• QAM Quadrature amplitude modulation
• the principle is to demodulate the signal received in one or more flexible bit (s) whose sign corresponds to the bit that would be provided by a hard decision detector and whose absolute value indicates the reliability of the decision of the demodulation module I / Q (in English "I / Q demapping module").
• the invention finds particular application in the field of telecommunications, for communication systems using the OFDM-CDMA technique, such as for example the Multi-Band OFDM Alliance system.
• an information bit stream d (k) is generated for each user k.
• Each bitstream is then coded using a channel coder 1O 1 , IO 2,..., 10 k, where K is the total number of users.
• the types of channel coders that can benefit from the present invention are all those whose corresponding decoder at the level of the receiver uses flexible input values.
• the block performing the channel coding can be constituted for example:
• the punching device makes it possible to obtain the coding efficiency desired for the application while the interleaver avoids the error packets on reception), or
• turbo encoder a punch and a bit interleaver
• an encoder using block codes with soft-input decoder of the LDPC (Low-Density Payty-Check Code) type.
• the coded bitstream b (k) is then transformed into a complex sequence of symbols MAQ-M (in English M-QAM) a (k), by means of a quadrature amplitude modulator 12i, 122, • -, 12k-
• M is the number of complex symbols or points in the constellation associated with the modulation.
• M is a power of 2
• M 2 m with m strictly positive integer.
• 2-QAM modulation also called BPSK
• QAM-4 modulation also called QPSK
• 8-QAM modulation and so on.
• m 2u with u strictly positive integer, we speak of square MAQ constellation (for example QAM-4, MAQ-16, MAQ-64, etc.). Since QAM-2 modulation (BPSK) only uses the in-phase channel (channel I), the formulas developed in the remainder of the description for the quadrature channel (Q-channel) are not to be taken into account for a single channel. QAM-2.
• BPSK QAM-2 modulation
• the symbol a (k) a ⁇ (k) + ja Q (k) corresponds to the complex symbol MAQ-M of the user k and ⁇ b
• the index q corresponds to the qth bit of the phase part of the signal (channel I) and the qth bit of the quadrature part of the signal (channel Q), and ml and mQ are the number of bits in channel I and on the Q path, respectively.
• the MAQ-M symbols are then spread in a module 14-i, 14 2 , ..., 14 K by spreading codes ⁇ c k] ⁇ ⁇ , 1 ⁇ k ⁇ K and 1 ⁇ £ ⁇ N (N being the number of sub-carriers) specific to each user k, then summed in a module 16. They are then subjected to a series-parallel conversion in a module 18, then a reverse fast Fourier transformation (TFRI, in English IFFT, "Inverse Fast Fourier Transform”) in a module 20. It is assumed that the transmitter separates the OFDM-CDMA symbols by a guard interval (IG) long enough to eliminate intersymbol interference (IES). This guard interval can be either cyclic or null prefix, both techniques allowing simple scalar equalization in the frequency domain.
• IG guard interval
• IES intersymbol interference
• Figure 2 shows the structure of a conventional receiver corresponding to the transmitter shown in Figure 1.
• the signal received from the transmission channel is first demodulated using a module 22 which synchronizes between the clocks of the transmitter and the receiver, eliminates the guard interval and applies a Fourier transformation.
• fast FFT, "Fast Fourier Transform"
• n the noise vector of dimension N * 1, N being the number of subcarriers, n b containing noise samples iid (independent and identically distributed) gaussian complex of variance ⁇ 2 , H is the diagonal matrix of dimension NxN representing the channel where the diagonal coefficient
• the frequency response of the channel is estimated for each subcarrier, in a module 24.
• These estimates H are used together with the spreading codes in order to achieve a linear equalization of the symbols received, in a module 26.
• the purpose of the mono- or multi-user equalizer is to reshape the received signal so that it corresponds as closely as possible to the points of the reference constellation. Nevertheless, in the presence of noise (thermal and multiple access), the points found do not coincide exactly with the initial constellation ⁇ . This is why, after the equalization, a soft I / Q demodulation is carried out in a module 28 before decoding the signal in a channel decoding module 30.
• the I / Q demodulation operation consists in finding the binary values from the complex symbols coming from the linear detector.
• the optimal values to be injected into the channel decoder are soft values, ie not directly the hard values "0" and "1".
• the soft I / Q demodulation will consist of calculating 4 values.
• the optimal soft (or metric) values to be injected into the channel decoder correspond to a Likelihood Ratio Logarithm (LRV).
• the constellation MAQ is divided into two complementary partitions of complex symbols, respectively S ⁇ containing the symbols with a "0" at position (l, q) and Sf 1 J containing the symbols with a
• the LRV can be expressed as follows:
• equation (4) the numerator of the logarithm summarizes the probabilities for all symbols with a "1" for the bit at position q and the denominator sumates the probabilities for all symbols with a "0" for the bit at the position q. These probabilities are decreasing exponential functions of the Euclidean distance between the received symbols and the reference symbols ⁇ .
• the decoding process which performs the dual operations of those performed by the encoder on transmission. If bit punching and interleaving operations are present on transmission, which is the case if convolutional codes or turbo-codes are used, the bit deinterleaving and then the de-interleaving operations are therefore performed after the demodulation I / Q flexible.
• Channel decoding finally makes it possible to retrieve the binary data transmitted. For example, in the case where a convolutional encoder is used on transmission, channel decoding uses the branch metric calculation, which uses the LRV given by equation (4). The branch metric for state s (i) of path i at time t is written:
• the i th complex data symbol received for the user k after a single-user detection is expressed in the following form, considering that the data has been transmitted and spread over N sub-carriers:
• Equation (8) corresponds to the case where all the users have an identical power and the equation (9) corresponds to the generalization in the case where the users have different powers.
• h is the estimate of h from the channel estimation module 24.
• Equation (13) For QAM-4, for large spreading factors, the calculation in equation (13) can be simplified as follows:
• Equation (14) For low spreading factors, simplification by equation (14) gives significantly worse results than for large spreading factors, because the central limit theorem is no longer checked.
• equation (15) is preferable to equation (13) when using a single-user MMSE equalizer, since computational complexity is reduced without significant loss of performance.
• equation (15) gives significantly worse results than equation (13) for the MRC (Maximum Ratio Combining) and ZF (zero forcing) equalizers. in English “Zero Forcing”), but not for the equalizer EGC (Equal Gain Combining).
• , q is half the distance between the boundaries of the partition relative to b
• , q is half the distance between the boundaries of the partition relative to b
• , 2 2 since the distance between the two borders is 4.
• the object of the invention is to overcome the disadvantages of the prior art, by optimizing the metric to be injected at the input of a channel decoder in an OFDM-CDMA system using non-binary orthogonal spreading codes.
• the present invention proposes a method of flexible data demodulation modulated according to a quadrature amplitude modulation of order greater than or equal to 4, in a communication system implementing a distributed multiple access technique by multi-carrier codes or OFDM-CDMA 1 using non-binary spreading codes, remarkable in that it comprises the steps of determining:
• P k is a parameter representative of the power applied to the k th user
• N is the number of subcarriers
• • c M is the value of the spreading code for the eighth subcarrier and the user k, • h, is the coefficient estimate of the transmission channel for the eighth subcarrier,
• Y k1 Q corresponds to the imaginary part of the complex symbol after equalization and despreading intended to be demodulated.
• the simplified metric proposed is particularly appropriate in the case of short spreading factors, which gives a sub-optimal decoding but which guarantees a low implementation complexity for a negligible loss of performance.
• the present invention finds a preferred application within the framework of the standard proposed by the consortium MBOA (Multi-Band OFDM Alliance).
• the logarithm of the likelihood ratio of the bits of the in-phase and quadrature channels for the bits bm ( ⁇ ), b m ( n) + 5o, b m (n) +1 , b m ( n) +5 i is given by the following equations:
• • lî ⁇ and h n + 50 are the estimates of the values of the frequency response of the transmission channel on the two sub-carriers.
• the present invention also proposes a device for flexible data demodulation modulated according to a quadrature amplitude modulation of order greater than or equal to 4, in a communication system implementing a multi-carrier code division multiple access technique or OFDM-CDMA, using non-binary spreading codes, notable in that it comprises a module for determining:
• P ⁇ ⁇ is a parameter representative of the power applied to the k th user
• N is the number of subcarriers
• c M is the spreading code of the value for the / -th sub-carrier and user k
• niQ. q is half the distance between partition boundaries relative to bQ , q and where:
• the invention also relates to a receiver adapted to implement a method as above.
• the invention also relates to a receiver comprising a device as above.
• FIG. 1 already described, schematically represents an OFDM-CDMA transmitter with channel coding of the conventional type
• FIG. 2 already described, schematically represents a receiver of conventional type corresponding to a transmitter of the type illustrated in FIG. 1;
• FIG. 3 schematically shows a part of a receiver capable of implementing a method according to the present invention, in a particular embodiment.
• the logarithm of the likelihood ratio makes it possible to pass complex symbols (I-channel and Q-channel) coming from an equalizer, such as the equalization / despreading module 32 shown in FIG. real (one per bit) indicating the reliability of the received bit before entering a channel decoder such as the channel decoder 36.
• the LRV thus realizes what is called a flexible I / Q demodulation operation 340, in a flexible I / Q demodulation module 34.
• This demodulation operation takes place between the equalization / despreading and channel decoding processes.
• the process of flexible I / Q demodulation, or calculation of the soft decisions at the input of the decoder uses the data resulting from the channel estimation, carried out by a channel estimation module 38, and the equalization performed by the module 32.
• the output of the soft I / Q demodulation module 34 corresponds to the LRV injected at the input of the channel decoder 36.
• the complex coefficients h f of the channel affecting the data symbols a k can be considered independent.
• the complex multiple access (IAM) and noise interference terms can be approximated by additive noise.
• Gaussian complex accordinging to the central limit theorem of zero mean and variance:
• Equation (20) corresponds to the case where all the users have an identical power and the equation (21) corresponds to the generalization in the case where the users have different powers.
• ⁇ is a reference symbol of the constellation associated with quadrature amplitude modulation
• P k is a parameter representative of the power applied to the k th user
• N is the number of subcarriers
• T is a subscript (positive integer) representative of the subcarrier (1 ⁇ £ ⁇ N)
• G • M is the value of the linear equalizer coefficient associated with £] th subcarrier and the user k
• S ⁇ and S ⁇ q are two complementary partitions of complex symbols respectively containing the symbols of the constellation with a "0" at the position (Q, q) and the symbols of the constellation with a "1" at the position (Q , q).
• y k, ⁇ corresponds to the real part of the complex symbol received after equalization and despreading.
• the imaginary part namely: where y k , Q corresponds to the imaginary part of the complex symbol received after equalization and despreading.
• the present invention also provides a simplified expression of the LRVs of an OFDM-CDMA system.
• formula (26) is preferable to equation (25) when using a single-user MMSE equalizer, since computational complexity is reduced without significant loss of performance.
• the input bits are first transformed into bipolar symbols, as follows:
• this modulation as 2 symbols a n and a ' n coming from a 4-QAM modulation and then used for non-binary code CDMA spread of length 2, by rewriting the equation (30) of the following way:
• h n and h n + 50 be the estimates of the values of the frequency response of the channel on the 2 subcarriers respectively modulated by S n and s n + 50 and are respectively g ⁇ and g n + 5 o the 2 coefficients of equalization employees.

Landscapes

• Engineering & Computer Science (AREA)
• Signal Processing (AREA)
• Computer Networks & Wireless Communication (AREA)
• Power Engineering (AREA)
• Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
• Error Detection And Correction (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=36926833

Family Applications (1)

Country Status (5)

Families Citing this family (16)

Family Cites Families (5)

• 2005
  • 2005-11-16 FR FR0511613A patent/FR2893433B1/fr not_active Expired - Fee Related
• 2006
  • 2006-11-14 EP EP06831117A patent/EP1949629A1/de not_active Withdrawn
  • 2006-11-14 US US12/093,925 patent/US7974357B2/en not_active Expired - Fee Related
  • 2006-11-14 JP JP2008540651A patent/JP5154431B2/ja not_active Expired - Fee Related
  • 2006-11-14 WO PCT/FR2006/002519 patent/WO2007057558A1/fr active Application Filing

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events